Cancer is both a disease of uncontrolled growth and loss of organ architecture. Indeed, the latter is how pathologists diagnose tumors and determine their 'grade'. Remarkably, as shown by animal studies and 3D cell culture models, enforcing a 'normal' context or architecture on malignant cells can cause them to behave like non-malignant cells, despite retaining a host of genetic mutations. This 'reversion' of malignant cells to a non-malignant phenotype is an example of the plasticity (malleability) of cells, which can change their organization and function in response to external signals in their microenvironment. Understanding how the interactions between a cell and its microenvironment control tissue phenotype and function has broad implications for understanding cancer and may ultimately lead to microenvironment-targeted therapies for certain malignancies. For decades some of us have been at the forefront of the role of the microenvironment in tissue-specificity and tumor biology and the role of tissue architecture in tumor cell induction, progression, and reversion. However, we still do not understand the physical principles that govern these processes. Indeed, we do not understand how we can make so many organs with such diverse characteristics when all the cells of these organs have the same genetic material. Nevertheless, our efforts have produced several versatile and robust models to study how organs maintain structure and how they lose it as cells progress to malignancy. In the process, we have discovered about 10 distinct agents that are able to revert a malignant or premalignant tissue to a phenotypically normal tissue, or. as in the case of mechanical shear stress, reduce the likelihood of reversion. All these agents represent potential avenues towards in situ reversion strategies and lend themselves to testing and modeling physical hypotheses. Physicists, mathematicians, and engineers at UCB, LBNL and UCSF have assembled to utilize state of the art tools for: understanding cellular plasticity and reversion research (e.g. designer 3D culture models' gene expression data with reverting agents; Bissell Lab), imaging and manipulation (e.g., optical and magnetic tweezers, PALM; Liphardt Lab), mechanobiology (e.g., AFM, live cell compression system; Fletcher Lab), and complementary modeling approaches (discreet continuum and Navier Stokes-based methods, agent-based modeling, and signaling network analysis; Sethian Lab, Costes, and Wolf, respectively), to uncover fundamentals of tissue plasticity and outline new strategies to understand both how tumors lose structure and how to restore tumors to a non-malignant state.